Academia.eduAcademia.edu

Figure 1 – uploaded by Marina Potapova

See full PDFdownloadDownload figure
List of metrics based on relative abundance of diatom indicator species

Table 1 List of metrics based on relative abundance of diatom indicator species

Related Figures (17)

To allow comparisons of the discriminative ability of diatom metrics based on datasets of different spatial extent, the national-scale NAWQA dataset was sub- divided into smaller-scale (“regional”) datasets. The main goal in delineating the individual regional datasets was to limit environmental and floristic variation in the data. The EPA “‘nutrient” ecoregions were used as basic spatial units. ““Nutrient’”’ ecoregions are aggregations of Omernik’s level three ecoregions (Omernik, 1995) suggested by the EPA as a spatial framework to investigate impacts of nutrient enrichment on fresh- water ecosystems throughout the U.S. (http://www.e- pa.gov/waterscience/criteria/nutrient/ecoregions). All 14 EPA “nutrient” ecoregions could not be used as spatial units, however, because the number of sites available in some ecoregions was limited. Therefore, some “nutrient” ecoregions were combined to obtain sufficiently large datasets. To do this, non-metric multidimensional scaling (NMS) was used to identify which “‘nutrient’”’ ecoregions were relatively similar in diatom species composition. Those which had a low number of sites, were geographically adjacent, and were located relatively close to each other in the ordination diagrams were combined. Two ordinations were used to make decisions about combining “nutrient” ecoregions: one with the dataset that included all 1240 sampling sites (one randomly selected Fig. 1. Map showing 1240 sampling sites and aggregations of ecoregions into five groups based on diatom assemblage similarity, geographic proximity, and number of diatom samples available for analysis. Several approaches were used to determine which diatom species were best indicators of nutrients. An indicator species analysis (Dufréne and Legendre, 1997) was carried out to identify which species were associated with the most nutrient-poor and the most nutrient-rich sites. This method searches for species, which not only have the highest specificity (mean relative abundance), but also the highest fidelity (frequency of occurrence) to a certain group of samples. All diatom samples that had corresponding TP < 10 pg L! were designated as ‘“low-TP” samples, those with TP > 100 wg L7! as “high-TP” sites, those with TN < 0.2 mg L7! as “low-TN” included all 1240 sampling sites (one randomly selected
To allow comparisons of the discriminative ability of diatom metrics based on datasets of different spatial extent, the national-scale NAWQA dataset was sub- divided into smaller-scale (“regional”) datasets. The main goal in delineating the individual regional datasets was to limit environmental and floristic variation in the data. The EPA “‘nutrient” ecoregions were used as basic spatial units. ““Nutrient’”’ ecoregions are aggregations of Omernik’s level three ecoregions (Omernik, 1995) suggested by the EPA as a spatial framework to investigate impacts of nutrient enrichment on fresh- water ecosystems throughout the U.S. (http://www.e- pa.gov/waterscience/criteria/nutrient/ecoregions). All 14 EPA “nutrient” ecoregions could not be used as spatial units, however, because the number of sites available in some ecoregions was limited. Therefore, some “nutrient” ecoregions were combined to obtain sufficiently large datasets. To do this, non-metric multidimensional scaling (NMS) was used to identify which “‘nutrient’”’ ecoregions were relatively similar in diatom species composition. Those which had a low number of sites, were geographically adjacent, and were located relatively close to each other in the ordination diagrams were combined. Two ordinations were used to make decisions about combining “nutrient” ecoregions: one with the dataset that included all 1240 sampling sites (one randomly selected Fig. 1. Map showing 1240 sampling sites and aggregations of ecoregions into five groups based on diatom assemblage similarity, geographic proximity, and number of diatom samples available for analysis. Several approaches were used to determine which diatom species were best indicators of nutrients. An indicator species analysis (Dufréne and Legendre, 1997) was carried out to identify which species were associated with the most nutrient-poor and the most nutrient-rich sites. This method searches for species, which not only have the highest specificity (mean relative abundance), but also the highest fidelity (frequency of occurrence) to a certain group of samples. All diatom samples that had corresponding TP < 10 pg L! were designated as ‘“low-TP” samples, those with TP > 100 wg L7! as “high-TP” sites, those with TN < 0.2 mg L7! as “low-TN” included all 1240 sampling sites (one randomly selected
Fig. 2. NMS ordinations of 428 “reference” sites (a) and of all 1240 sites (b). Diagrams show positions of ‘‘nutrient” ecoregion centroids and their aggregation into five groups based on criteria described in the text. “Nutrient” ecoregions are: (1) Willamette and Central Valleys, (2) Western Forested Mountains, (3) Xeric West, (4) Great Plains Grass and Shrublands, (5) South Central Cultivated Great Plains, (6) Corn Belt and Northern Great Plains, (7) Mostly Gla- ciated Dairy Region, (8) Nutrient Poor Largely Glaciated Upper Midwest and Northeast, (9) Southeastern Temperate Forested Plains and Hills, (10) Texas—Louisiana Coastal and Mississippi Alluvial Plains, (11) Central and Eastern Forested Uplands, (12) Southern Coastal Plain, (13) Southern Florida Coastal Plain, and (14) Eastern Coastal Plain. No “reference” sites were available in “‘nutrient” ecoregion 10. NMS ordinations (Fig. 2) showed that five ecoregion groups varied in their relative homogene- ity. For instance, ecoregions forming the group “Eastern Plains’ appeared to be more diverse than other groups because their centroids were located further apart than ecoregion centroids in other groups. This heterogeneity may be partly a function of the different number of sites representing each
Fig. 2. NMS ordinations of 428 “reference” sites (a) and of all 1240 sites (b). Diagrams show positions of ‘‘nutrient” ecoregion centroids and their aggregation into five groups based on criteria described in the text. “Nutrient” ecoregions are: (1) Willamette and Central Valleys, (2) Western Forested Mountains, (3) Xeric West, (4) Great Plains Grass and Shrublands, (5) South Central Cultivated Great Plains, (6) Corn Belt and Northern Great Plains, (7) Mostly Gla- ciated Dairy Region, (8) Nutrient Poor Largely Glaciated Upper Midwest and Northeast, (9) Southeastern Temperate Forested Plains and Hills, (10) Texas—Louisiana Coastal and Mississippi Alluvial Plains, (11) Central and Eastern Forested Uplands, (12) Southern Coastal Plain, (13) Southern Florida Coastal Plain, and (14) Eastern Coastal Plain. No “reference” sites were available in “‘nutrient” ecoregion 10. NMS ordinations (Fig. 2) showed that five ecoregion groups varied in their relative homogene- ity. For instance, ecoregions forming the group “Eastern Plains’ appeared to be more diverse than other groups because their centroids were located further apart than ecoregion centroids in other groups. This heterogeneity may be partly a function of the different number of sites representing each
M. Potapova, D.F. Charles/ Ecological Indicators 7 (2007) 48-70 Fig. 3. Values of trophic diatom metrics for 209 sites in the “Western Mountains” ecoregion grouped into 3 TP categories. Metric abbreviations are given in Table 1. The gray bars in each box plot show the 25th and 75th percentiles of the data, the black bar inside the gray bar shows the median, the whiskers mark 1.5 x interquartile range, circles show outliers between 1.5 x and 3.0x interquartile range, and stars show extreme values outside 3.0x interquartile range.
M. Potapova, D.F. Charles/ Ecological Indicators 7 (2007) 48-70 Fig. 3. Values of trophic diatom metrics for 209 sites in the “Western Mountains” ecoregion grouped into 3 TP categories. Metric abbreviations are given in Table 1. The gray bars in each box plot show the 25th and 75th percentiles of the data, the black bar inside the gray bar shows the median, the whiskers mark 1.5 x interquartile range, circles show outliers between 1.5 x and 3.0x interquartile range, and stars show extreme values outside 3.0x interquartile range.
M. Potapova, D.F. Charles/ Ecological Indicators 7 (2007) 48-70 Fig. 4. Values of trophic diatom metrics for 209 sites in the “Western Mountains” ecoregion grouped into 3 TN categories. Metric abbreviations are given in Table 1. Box plot symbols as in Fig. 3.Fig. 4. Theoretical graphs and data plots of (a) u«/Up, (b) Zo/h, and (c) do/h vs. A for MODIS/ IGBP land cover type grasslands.
M. Potapova, D.F. Charles/ Ecological Indicators 7 (2007) 48-70 Fig. 4. Values of trophic diatom metrics for 209 sites in the “Western Mountains” ecoregion grouped into 3 TN categories. Metric abbreviations are given in Table 1. Box plot symbols as in Fig. 3.Fig. 4. Theoretical graphs and data plots of (a) u«/Up, (b) Zo/h, and (c) do/h vs. A for MODIS/ IGBP land cover type grasslands.
M. Potapova, D.F. Charles/ Ecological Indicators 7 (2007) 48-70 Fig. 5. Values of trophic diatom metrics for 138 sites in the ‘Eastern Highlands” grouped into 3 TP categories. Metric abbreviations are given in Table 1. Box plot symbols as in Fig. 3.
M. Potapova, D.F. Charles/ Ecological Indicators 7 (2007) 48-70 Fig. 5. Values of trophic diatom metrics for 138 sites in the ‘Eastern Highlands” grouped into 3 TP categories. Metric abbreviations are given in Table 1. Box plot symbols as in Fig. 3.
M. Potapova, D.F. Charles/ Ecological Indicators 7 (2007) 48-70 Fig. 6. Values of trophic diatom metrics for 138 sites in the “Eastern Highlands” grouped into 3 TN categories. Metric abbreviations are given in Table 1. Box plot symbols as in Fig. 3.
M. Potapova, D.F. Charles/ Ecological Indicators 7 (2007) 48-70 Fig. 6. Values of trophic diatom metrics for 138 sites in the “Eastern Highlands” grouped into 3 TN categories. Metric abbreviations are given in Table 1. Box plot symbols as in Fig. 3.
Spearman correlation coefficients between diatom metrics (abbreviations in Table 1) and nutrient (TP or TN) concentrations in the five regional NAWQA datasets Table 2 and ecological studies allow recognition of taxa with good indicator properties (Round, 2004). their own numerous observations made in The Netherlands, to assign diatom taxa to ecological categories. Correlations between metrics or indices and nutrient concentrations in test sites located in the same geographic areas for which metrics were developed were strong. For example, the correlation coefficient between the Austrian Trophic Index and TP in Austria was 0.85 (Rott et al., 2003). The correlation between the DAIpo index and TP in Japanese streams was approximately 0.70-0.80 (Watanabe et al., 1988). We think that better coverage (more sites per area) would further increase dis- criminative abilities of our metrics because autecology of more diatom species could be quantified.
Spearman correlation coefficients between diatom metrics (abbreviations in Table 1) and nutrient (TP or TN) concentrations in the five regional NAWQA datasets Table 2 and ecological studies allow recognition of taxa with good indicator properties (Round, 2004). their own numerous observations made in The Netherlands, to assign diatom taxa to ecological categories. Correlations between metrics or indices and nutrient concentrations in test sites located in the same geographic areas for which metrics were developed were strong. For example, the correlation coefficient between the Austrian Trophic Index and TP in Austria was 0.85 (Rott et al., 2003). The correlation between the DAIpo index and TP in Japanese streams was approximately 0.70-0.80 (Watanabe et al., 1988). We think that better coverage (more sites per area) would further increase dis- criminative abilities of our metrics because autecology of more diatom species could be quantified.
Fig. 7. Values of trophic diatom metrics for 397 Mid-Atlantic stream sites (MAHA dataset) grouped into 3 TP and 3 TN categories. Metric abbreviations are given in Table 1. Box plot symbols as in Fig. 3.
Fig. 7. Values of trophic diatom metrics for 397 Mid-Atlantic stream sites (MAHA dataset) grouped into 3 TP and 3 TN categories. Metric abbreviations are given in Table 1. Box plot symbols as in Fig. 3.
Diatoms indicating low (—) or high (+) nutrients in U.S. rivers. Species were included in this list as a result of two analyses. First, indicator species analysis (Dufréne and Legendre, 1997) determined species indicator values in groups of samples with corresponding TP < 10 wg L~' (low-TP), TP > 100 pg L~ (high-TP), TN < 0.2 mg L! (low-TN), and TN > 3.0 mg L (high-TN). Diatoms with indicator values greater than 5 (P < 0.05) are listed here, and their indicator values are shown. The second analysis was based on calculation of species abundance- weighted means. Diatoms with TP and TN abundance-weighted means above the 75th percentile or below the 25th percentile of all values are listed here and marked by asterisks.
Diatoms indicating low (—) or high (+) nutrients in U.S. rivers. Species were included in this list as a result of two analyses. First, indicator species analysis (Dufréne and Legendre, 1997) determined species indicator values in groups of samples with corresponding TP < 10 wg L~' (low-TP), TP > 100 pg L~ (high-TP), TN < 0.2 mg L! (low-TN), and TN > 3.0 mg L (high-TN). Diatoms with indicator values greater than 5 (P < 0.05) are listed here, and their indicator values are shown. The second analysis was based on calculation of species abundance- weighted means. Diatoms with TP and TN abundance-weighted means above the 75th percentile or below the 25th percentile of all values are listed here and marked by asterisks.
Appendix A. (Continued )
Appendix A. (Continued )
M. Potapova, D.F. Charles/ Ecological Indicators 7 (2007) 48-70
M. Potapova, D.F. Charles/ Ecological Indicators 7 (2007) 48-70
Appendix A. (Continued )
Appendix A. (Continued )
Appendix A. (Continued )
Appendix A. (Continued )
Related topics:
Ecological IndicatorsIndicator speciesSimilarity Metric
580 California St., Suite 400
San Francisco, CA, 94104
© 2025 Academia. All rights reserved